Hi Matthew,
That appears to have worked-- I fed it the FEAT-generated design matrix
and the filtered_func_data as inputs. Can you direct me to any
documentation for fsl_glm? I can't seem to find anything about it on the
FSL website. It must be just a simple GLM I'm assuming, right? I just
want to make sure I know exactly what is going on for this analysis,
like any pre-processing steps or anything out of the ordinary. Thanks
again!
Andrew Dumas
On Mon, 2010-10-04 at 10:09 +0100, Matthew Webster wrote:
> Hi,
> You're right - FILM does some additional processing on the input data as well as fitting the model, can you try using fsl_glm to fit the full model and then use the residuals from that to obtain the full model from the data?
>
> Many Regards
>
> Matthew
>
> > Hi Matthew,
> > Actually, that was the first thing I tried and it didn't work
> > unfortunately. Is there some processing between the filtered_func_data
> > and the residuals that I can do on the filtered_func_data? Thanks again,
> >
> > Andrew
> >
> > On Fri, 2010-10-01 at 15:27 +0100, Matthew Webster wrote:
> >> Hello Andrew,
> >> If you subtract the stats/res4d file from the filtered_func_data in your FEAT directory, you should get a 4D volume containing the full-model.
> >>
> >> Many Regards
> >>
> >> Matthew
> >>> Hi,
> >>>
> >>> Is there any built-in way to generate a 4D volume of the full model, following output from FEAT? Much like in FSLVIEW, when viewing a time course, you can overlay the full model for a particular voxel in the plot window-- I'd like to do that for the whole volume so that I can obtain the full model at each voxel using MATLAB. I know featquery can do this on a limited basis-- i.e. output the time course and model for a particular voxel or masked region as a text file, but is there a way to output as Nifti 4D volume?
> >>> I'm using FLOBS with basis functions and I'm trying to look at some timing parameters like time-to-peak, and return to baseline in multiple ROIs, with further output to Free Surfer, so it would be extremely useful to be able to do this. I'd hate to have to put this together manually-- using the PEs and design matrix.
> >>>
> >>> Thanks everyone in advance!
> >>>
> >>> Andrew Dumas
> >>> [log in to unmask]
> >>> Research Assistant
> >>> MGH Neurology
> >>> Martinos Center for Biomedical Imaging
> >>>
> >>
> >>
> >
>
>
|